/**
 * Embedding backend interface that all embedding providers must implement.
 * Supports both single text and batch embedding operations.
 */
export interface EmbeddingBackend {
    /** Display name of the embedding backend (e.g., 'gemini', 'ollama') */
    name: string;
    /**
     * Initialize the backend, validating credentials and connectivity.
     * @throws Error if initialization fails (e.g., invalid API key, server unreachable)
     */
    initialize(): Promise<void>;
    /**
     * Generate an embedding vector for a single text.
     * @param text - The text to embed
     * @returns A promise resolving to the embedding vector (array of numbers)
     */
    embed(text: string): Promise<number[]>;
    /**
     * Generate embedding vectors for multiple texts in a single batch.
     * More efficient than calling embed() multiple times.
     * @param texts - Array of texts to embed
     * @returns A promise resolving to an array of embedding vectors
     */
    embedBatch(texts: string[]): Promise<number[][]>;
    /**
     * Get the dimensionality of the embedding vectors produced by this backend.
     * @returns The number of dimensions in the embedding vectors
     */
    getDimensions(): number;
    /**
     * Get the model identifier used by this backend.
     * @returns The model name/identifier (e.g., 'nomic-embed-text', 'gemini-embedding-001')
     */
    getModel(): string;
}
/**
 * Configuration options for embedding backends.
 */
export interface EmbeddingConfig {
    /** Which embedding backend to use */
    backend: 'ollama' | 'gemini' | 'local';
    /** Model name/identifier (backend-specific) */
    model?: string;
    /** API key for cloud-based backends (Gemini) */
    apiKey?: string;
    /** Base URL for the embedding API (useful for Ollama or custom endpoints) */
    baseUrl?: string;
    /** Rate limiting: maximum requests per second */
    rateLimitRps?: number;
    /** Rate limiting: maximum burst capacity (default: 10) */
    rateLimitBurst?: number;
    /**
     * Maximum number of texts to process in a single batch request.
     * Large batches are automatically split into smaller chunks to prevent
     * timeouts, memory issues, and API rate limit errors.
     * Default: 100 for Ollama
     */
    batchSize?: number;
    /**
     * Maximum number of concurrent batch requests (Ollama only).
     * Controls how many batch requests are sent in parallel.
     * Default: 4 for Ollama
     */
    concurrency?: number;
}
/**
 * Split an array into chunks of specified size.
 * @param array - The array to chunk
 * @param size - Maximum size of each chunk
 * @returns Array of chunks
 */
export declare function chunkArray<T>(array: T[], size: number): T[][];
/**
 * Default embedding configuration using Google Gemini.
 * Gemini offers a free tier with 1500 RPM.
 * Get API key at: https://aistudio.google.com/app/apikey
 */
export declare const DEFAULT_CONFIG: EmbeddingConfig;
/**
 * Information about a backend fallback that occurred during initialization.
 * This is returned when the configured backend fails and we fall back to another.
 */
export interface BackendFallbackInfo {
    /** Whether a fallback occurred */
    occurred: true;
    /** The backend that was originally configured */
    originalBackend: string;
    /** The backend we fell back to */
    fallbackBackend: string;
    /** The error message from the original backend failure */
    reason: string;
}
/**
 * Result of creating an embedding backend, including optional fallback info.
 */
export interface CreateBackendResult {
    /** The initialized embedding backend */
    backend: EmbeddingBackend;
    /** Information about fallback if one occurred */
    fallback?: BackendFallbackInfo;
}
//# sourceMappingURL=types.d.ts.map